Reaction Time as a Digital Biomarker

Reaction time is how long a person takes to respond to a stimulus. In clinical research the measurement that matters is the psychomotor vigilance test, a simple visual reaction time task sustained over several minutes that is designed to detect lapses in attention rather than peak speed.

Its value comes from a specific property: performance on it degrades reliably with sleep loss, shows almost no learning effect across repeated administrations, and is sensitive to both total and partial sleep deprivation. That combination is unusual. Most cognitive tests improve with practice, which makes them awkward for repeated measurement, and the vigilance test's resistance to practice effects is exactly why it can be administered daily for weeks.

This page is written for that research context. General searches for reaction time are dominated by consumer testing and gaming, which is a different subject entirely.

The task presents a stimulus at random intervals and the participant responds as quickly as possible. The classic version runs for ten minutes; brief versions lasting three minutes were developed and validated for settings where the full test is impractical, including spaceflight, and adaptive duration versions now exist.

The outcome is not mean reaction time alone. The informative measures are the number of lapses, meaning responses slower than a threshold, the slowest ten percent of responses, and metrics that combine speed and lapse frequency. Mean reaction time on its own understates the effect of sleep loss because the degradation appears mostly in the tail of the distribution.

Delivery is now typically on a phone or tablet, which introduces device dependent timing that studies should control by keeping hardware consistent.

The dominant use is sleep and circadian research, where the test quantifies the functional consequence of sleep restriction, shift work and circadian misalignment. It is also used in fatigue risk management research, in studies of sedating medication where impairment is a safety question, and as a repeated cognitive measure in neurological and psychiatric trials.

Because it can be administered daily without meaningful practice effects, it suits designs that need a trajectory rather than two distant points. It is reported alongside cognitive instruments such as the Trail Making Test and the Frontal Assessment Battery, which assess broader executive function, and alongside subjective measures, because people are consistently poor at judging their own impairment from sleep loss.

The test measures sustained attention, which is one narrow cognitive domain. It is not a measure of general cognition, memory or executive function, and treating a change in it as evidence of cognitive decline overreaches.

Results depend on the hardware and software delivering it, since millisecond timing differs between devices and browsers, so absolute values are not comparable across platforms. Motivation, time of day and caffeine all affect performance independently of the condition under study.

Finally, the version matters. Ten minute, three minute and adaptive versions produce different values with different sensitivity, and a study reporting reaction time without specifying which version was used has not reported enough to interpret.

• Basner M, et al. Validity and sensitivity of a brief psychomotor vigilance test (PVT-B) to total and partial sleep deprivation. Acta Astronaut. 2011. pubmed.ncbi.nlm.nih.gov
• Lim J, Dinges DF. Sleep deprivation and vigilant attention. Ann N Y Acad Sci. 2008. pubmed.ncbi.nlm.nih.gov
• Basner M, et al. A new likelihood ratio metric for the psychomotor vigilance test and its sensitivity to sleep loss. J Sleep Res. 2015. pubmed.ncbi.nlm.nih.gov
• Basner M. Ultra-short objective alertness assessment: an adaptive duration version of the 3 minute PVT. Sleep Adv. 2022. pubmed.ncbi.nlm.nih.gov